Comparison of assay methods for quantifying sex hormone concentrations across the menstrual cycle in rhesus macaques†

Abstract

Immunoassays have been the preferred method for steroid hormone analysis for more than 50 years. Automated immunoassays (AIAs) offer high throughput, rapid data turnaround, and low cost for measuring steroid hormone concentrations. The application of liquid chromatography-tandem mass spectrometry (LC-MS/MS) for steroid quantification provides greater specificity and selectivity for individual steroids, the ability to simultaneously analyze multiple steroids, and high throughput and automation. We compared AIA and LC-MS/MS for analysis of 17beta-estradiol (E2) and progesterone (P4) over the course of several menstrual cycles in 12 rhesus macaques (Macaca mulatta). Serum samples were collected every 4 days across four menstrual cycles from each monkey. AIAs were performed on a Roche cobas e411 analyzer. LC-MS/MS analysis was performed on a Shimadzu-Nexera-LCMS-8060 instrument. Scatter plots with Passing-Bablok regression showed excellent agreement between AIA and LC-MS/MS for both E2 and P4. Bland-Altman plots revealed no bias for either method; however, AIA overestimated E2 at concentrations >140 pg/ml and underestimated P4 at concentrations >4 ng/ml compared to LC-MS/MS. A comparison of testosterone concentrations measured by AIA and LC-MS/MS in the same samples was also performed. In contrast to E2 and P4, AIA and LC-MS/MS yielded significantly different results for testosterone concentrations, with AIA consistently underestimating concentrations relative to those obtained by LC-MS/MS. Well-characterized automated immunoassays are an excellent tool for daily monitoring of monkey menstrual cycles or providing single data points requiring fast turnaround. In certain situations where AIAs may provide inaccurate estimations of E2 and P4 concentrations, LC-MS/MS assays are preferable.

Keywords: LC-MS/MS; immunoassay; menstrual cycle; nonhuman primate; steroid hormone.

Figure 1

Scatter plot of estradiol (E2) measurements by automated immunoassay (Roche cobas e411) and LC-MS/MS (Shimadzu-Nexera-LCMS-8060). (A) Scatter plot showing all E2 data. The equal concentration (1:1) line (solid line), Passing–Bablok regression line (dotted line), and regression equation are shown on the graph. (B) scatter plot showing data where E2 concentrations are ≤140 pg/ml by LC-MS/MS. The equal concentration line (solid line) and Passing–Bablok regression line (dotted line) for the entire dataset is shown on the graph.

Figure 2

Scatter plot of progesterone (P4) measurements by automated immunoassay (Roche cobas e411) and LC-MS/MS (Shimadzu-Nexera-LCMS-8060). (A) Scatter plot showing all P4 data. The equal concentration (1:1) line (solid line), Passing–Bablok regression line (dotted line), and regression equation are shown on the graph. (B) Scatter plot showing data where P4 concentrations are ≤4 ng/ml by LC-MS/MS. The equal concentration line (solid line) and Passing–Bablok regression line (dotted line) for the entire dataset is shown on the graph. A scatter plot where P4 concentrations are ≤0.5 ng/ml by LC-MS/MS is included in the Supplementary Materials (Supplementary Figure S1).

Figure 3

Bland–Altman plot of estradiol concentrations measured by LC-MS/MS compared to estradiol concentrations measured by AIA. The solid line represents the mean difference of the assay methods and dashed lines represent 2 standard deviations above and below the mean difference (95% limits of agreement).

Figure 4

Bland–Altman plot of progesterone concentrations measured by LC-MS/MS compared to progesterone concentrations measured by AIA. The solid line represents the mean difference of the assay methods and dashed lines represent 2 standard deviations above and below the mean difference (95% limits of agreement).

Figure 5

Representative plot of sex hormone concentrations across four menstrual cycles in a female rhesus macaque (Animal 2; data from Supplementary Table S1). Estradiol and progesterone concentrations were measured by AIA on a Roche cobas e411 instrument and LC-MS/MS on a Shimadzu-Nexera-LCMS-8060 instrument.

Figure 6

Scatter plot of testosterone (T) measurements by automated immunoassay (Roche cobas e411) and LC-MS/MS (Shimadzu-Nexera-LCMS-8060). (A) Scatter plot showing all T data. The equal concentration (1:1) line (solid line), Passing–Bablok regression line (dotted line), and regression equation are shown on the graph. (B) Scatter plot showing data where T concentrations are ≤0.6 ng/ml by LC-MS/MS. The equal concentration line (solid line) and Passing–Bablok regression line (dotted line) for the entire dataset is shown on the graph.

Figure 7

Bland–Altman plot of testosterone concentrations measured by LC-MS/MS compared to testosterone concentrations measured by AIA. The solid line represents the mean difference of the assay methods and dashed lines represent 2 standard deviations above and below the mean difference (95% limits of agreement).